1. Field of the Invention
The present invention relates to an autonomous driver assistance system and a method therefor, and more particularly to an autonomous driver assistance system and an autonomous driving method thereof combining road detection, autonomous driving, telematics and vehicle safety.
2. Description of the Related Art
Current vehicles have been equipped with more and more driver assistance apparatuses, such as the early passive driving assistance apparatuses including ABS (Anti-Lock Braking System), air bag, EBD (Electronic Brake Force Distribution), and the like, to provide drivers the capability of controlling the vehicles or protecting drivers' safety upon vehicle collision in an emergency braking condition. However, such passive driving assistance apparatuses still fall short as far as driving safety is concerned. To further improve driving safety and vehicle safety, active safety perception apparatuses, such as LDWS (Lane Departure Warning System), FCWS (Forward Collision Warning System) and the like, have been developed in the market one after the other to alert drivers in advance when detecting any emergency condition in front of the vehicles. The LDWS constantly detects lane or sideline markings in front of vehicles. In the event of any unintentional departure from an original lane or road border, the LDWS will trigger a warning tone to alert drivers. The drawback of the LDWS is that lane and sideline markings must be clearly recognizable. During dusky, snowy or fogy condition, the LDWS may not normally function for failing to clearly recognize lane and sideline markings. The FCWS detects a distance to a vehicle ahead through imaging or radar detection. If the distance is too close, the FCWS immediately triggers an alert or activates the brake system of the vehicle to maintain a safe driving distance or conduct an emergency stop so as to prevent vehicle crash or mitigate impact force. The drawback of the FCWS is that place of collision can never be predicted and the vehicle ahead must enter a system detection zone before the FCWS normally operates (the FCWS does not work when other vehicles are on the lateral sides of the vehicle).
The active safety perception apparatuses can only alert drivers or do an emergency stop upon departure from lane or sideline marking or imminent collision. Basically, it is still the responsibility of drivers to drive the vehicles to destinations. When the active safety perception apparatuses are applied to autonomous driving, some deficiencies appear. For example, in addition to detection of obstructions and hazards on the road, autonomous driving further requires route planning and road condition recognition.
A recently developed active safety perception apparatus as disclosed in U.S. Pat. No. 8,195,394 entitled “Object detection and classification for autonomous vehicles” is to equip a vehicle with sensors for measuring distance and for imaging, 3D radar (Three-dimensional Light Radar) and GPS (Global Positioning System) maps. The sensors are used to detect moving objects and obstacles on the roadway. The 3D radar and the GPS maps are used to avoid stationary obstacles on the roadway by the comparison with actual scenes of roadways so as to enhance the recognition of neighboring obstacles. Although the 3D radar can accurately scan topography or obstacles, the cost of the 3D radar is rather costly. Besides, a gigantic database is also required to store map information beforehand. The cost of building this active safety perception apparatus is therefore prohibitive.
Another U.S. Pat. No. 7,065,245 entitled “Image processing apparatus and the method thereof” involves a method and a device for performing a stereo matching on a pair of color images constituted by a plurality of primary colors. This conventional technique employs image recognition means to recognize signals of direction indicators and predicts a moving direction of the vehicle. However, such technique fails to disclose how to effectively recognize the signals of the direction indicators and how to combine itself with autonomous driving technique or vehicle control technique of the vehicle, making such technique not well-feathered to be combined with autonomous driving applications.